Infrastructure-as-a-Service (IaaS) has changed the game.
Over a decade ago, Katy Perry was setting off fireworks, Adele was rolling in the deep and the Party Rockers were in the house that night… But what early adopters were also doing was moving their virtual servers to the cloud.
Because it had (and still continues to have) an amazing value proposition.
Fast-forward to 2022, and IaaS is still very popular as a long-term strategy for many organisations. It works and is massively lucrative for them. It’s also the reason why hyperscalers such as GCP, AWS and Azure put as much effort into their IaaS offerings as they do.
But is it really modernisation? Are you enabling your team? Are you offering better services to your customers, be it internal or external?
Whatever the answer, let us provide some perspective with a series on:
- The Good (Part 1);
- The Bad, & The Ugly (Part 2); and
- Ways to get the best out of your cloud-based VMs! (Part 3)
Before We Begin, What Is a Virtual Machine (VM)?
In the late 90s and early 2000s physical servers made original equipment manufacturers (OEMs) more money than they could count.
Fairly so, as well; It was reliable and outlived its warranty by decades.
The OEMs were ahead of the game. They knew they could get more silicon into the same form-factor which meant that they could charge more.
The problem was, there was no use for single-use appliances capable of more than was needed. Some free-thinking organisations realised that software could abstract the immense capabilities and split it into multiple virtualised instances utilising a reserved and increasingly well-orchestrated portion of the larger piece of hardware.
This spawned an industry of building high-capacity compute capabilities beyond what could have been imagined. Mostly for the purpose of segmenting them into individual systems that were capable of managing distributed workloads.
This silicon-based schizophrenia facilitated the building of massive compute farms with a reduced hardware footprint – while building VMs which would handle the workloads effortlessly and still maximise on the hardware investment. The next evolutionary step was to provide highly available and redundant systems based on this technology, which meant that we got geographically distributed systems which gave us maintainable scale at essentially half the cost.
This led to the ‘Software Defined’ world where the need for orchestration and manageability became key. We had more servers than engineers, so we needed to back them with the tools to manage more servers with less people.
This massive orchestration drives what we know as the cloud today. An excess of hardware married to a lack of human capital to manage it meant that software needed to do a lot more heavy lifting.
Organisations with either a large hardware budget or a surplus of unutilised hardware led to them building these environments where it could be used publicly.
So long story short, virtual servers are built on slices of physical hardware. With the tools we have these days and the optimization of hardware, we can deliver this without limitations in increasingly smaller facilities.
Why Are VMs In The Cloud So Awesome?
The reasons are plentiful.
Storage arrays for running VMs on your own hardware are ridiculously expensive.
As are modern compute mechanisms. In addition, you need specialist engineers to maintain and monitor the system, as well as a network to meet consumer demands. Before you know it, an army of engineers are keeping the lights on.
Moving it to the cloud abstracts the entire physical layer, including the upfront costs. The Operating system management is the only focus. All the rest, including capex costs for hardware, are no longer a concern.
The reduction in licensing complexity removes an even bigger headache.
In the cloud, the latest operating systems ensure constant evolution in the core OS stack. No more mad scrambles every 10 years to reduce risk stemming from deprecated operating systems.
Cloud providers help you monitor, maintain and auto-remediate.
This solves a quagmire of problems that historically had to be resolved with expensive 3rd party software – and which required a dedicated engineering team to manage.
Additional orchestration extends to scaling systems up and down based on demand to manage costs. We have written about this extensively on our blog.
You can streamline your teams and allow your business to focus on its most important activities…
… making VMs in the cloud make a very attractive value proposition. We do not advocate downsizing, but rather re-assigning those engineers (who previously focused on the nuts and bolts) to add value on your cloud journey.
As Great As They Are, VMs Aren’t All Sunshine…
VMs aren’t all sunshine and roses. In fact, they present challenges even under the best environments or circumstances.
While provisioning, licensing and general management come to mind, those aren’t exactly new problems.
In the next piece on VMs In The Cloud, I’ll unpack the Bad and The Ugly – a breakdown of VM sprawl, hyperscale lock-in, and the (heavy) costs that come with those cloud-based VMs.